Early studies focused on pathophysiology comparing gram positive and gram negative organisms, the role of endotoxemia, and the efficacy of anti-endotoxin therapies such as lipid A analogs and antibodies.[unreadable] [unreadable] Nitric oxide was examined as an important mediator of septic shock. Non-selective nitric oxide synthase inhibitors were sometimes toxic and never beneficial.[unreadable] [unreadable] Normal volunteers challenged with endotoxin were found to release increased amounts of nitric oxide. Although ibuprofen blocked endotoxin-induced increases in nitric oxide production, blood pressure was unaffected, suggesting that other mechanisms compensated to maintain vasodilation.[unreadable] [unreadable] Severity of illness (risk of death) was found to influence the therapeutic efficacy of anti-inflammatory agents in septic shock. This finding was used by the FDA to re-analyze the PROWESS trial of rhAPC (Xigris) and led to initial approval only for patients with a high risk of death. The lack of rhAPC efficacy in patients with a low risk of death was confirmed in the ADDRESS trial.[unreadable] [unreadable] The administration of L-arginine without or with N-acetylcysteine in a canine model of septic shock was found to be harmful. L-arginine is a common component of immunonutrition formulas that are marketed for use in critically ill patients.[unreadable] [unreadable] Intra-aortic balloon counterpulsation was found to prolong survival in a hypodynamic canine model of Staphylococcus aureas pneumonia induced septic shock.